openclaw - 💡(How to fix) Fix Heartbeat scheduler resets on gateway restart instead of maintaining wall-clock cadence [1 participants]

Over 7 days (Mar 26 – Apr 2, 2026), the heartbeat fired 1–3 times per day at best, almost always clustered within minutes of a gateway restart. On multiple days, zero heartbeats fired during the 07:00–23:00 active window.

TL;DR

To achieve a wall-clock-anchored heartbeat cadence, consider implementing a cron-style scheduling system that is not dependent on the gateway's restart time.

Guidance

• Review the current scheduler implementation to identify why it resets the 2-hour interval timer on every gateway restart, and consider using a persistent storage mechanism to maintain the schedule.
• Explore using a cron-style scheduling library or system that can handle wall-clock-anchored schedules, such as node-cron or cron-expression, to replace the current every-based scheduling.
• Consider adding a mechanism to handle the case where the gateway restarts during the inactive hours (23:00-07:00) to ensure the next heartbeat fires at the next scheduled time (07:00) rather than immediately after restart.
• Evaluate the feasibility of supporting explicit cron-style schedule expressions, such as at: ["07:00", "09:00", "11:00", ...], as an alternative to the every parameter.

Example

{
  "heartbeat": {
    "at": ["07:00", "09:00", "11:00", "13:00", "15:00", "17:00", "19:00", "21:00"],
    "activeHours": { "start": "07:00", "end": "23:00" },
    "model": "anthropic/claude-haiku-4-5",
    "target": "last",
    "lightContext": true
  }
}

Notes

The current implementation's reliance on the every parameter and lack of persistence may be contributing to the observed behavior. A cron-style scheduling system can provide more precise control over the schedule and ensure that the heartbeat cadence is maintained even after gateway restarts.

Recommendation

Apply a workaround by implementing a cron-style scheduling system, as it provides a more robust and reliable way to maintain a wall-clock-anchored heartbeat cadence. This approach can help mitigate the issues caused by the current scheduler's behavior and provide a more predictable heartbeat schedule.